Research On Target Recognition And Grab Technology Of Binocular Vision Robot

Machine vision is a fast development direction in the field of artificial intelligence.With the continuous advancement of machine vision technology,it has promoted the progress of industrial automation,intelligent security and artificial intelligence,a nd brought more potential and opportunity for developments in various fields.Among them,binocular vision is an important branch in the field of machine vision.Introducing vision into the robot's grasping,which is a combination of robotic control techno logy and machine vision,has a profound impact on improving the efficiency of industrial production.The subject mainly introduces binocular vision technology into traditional industrial teaching robots,enabling robots to use vision to realize the recognition,location and capture of target objects.The main research work is as follows:First,research on camera calibration technology,establish camera imaging related coordinate system and analyze the principle and calibration method of camera calibration.Based on the calibration principle of Zhang Zhengyou,camera calibration experiments were performed to obtain camera parameters.The calibration results were compared.The camera was calibrated using Opencv programming method to obtain camera parameters.Secondly,the image feature extraction and matching algorithm is studied.The images captured by the camera are often unable to directly extract and analyze the features due to noise and other factors.Therefore,before the feature extraction,the filtering algorithm of the image is studied.Experiments are carried out to compare the filtering effects of different filtering algorithms,and the median filtering algorithm is selected to preprocess the images.By analyzing different types of image features,it is proposed to extract the extreme point features of the image in the image space domain,and use the SIFT algorithm to extract and match the features.The feature extraction algorithm is optimized by RANSAC algorithm.The realization of visual capture requires identification and positioning of the target object,and the workpiece recognition is performed using the template recognition principle of extreme points.The object is subjected to recognition experiments under different working conditions.The object is positioned by means of centroid positioning,and the three-dimensional coordinates of the centroid of the object are calculated.Finally,the overall framework of the binocular visual capture system is constructed,the functions of each module are analyzed,and the camera,lens and end hand are selected and analyzed.The kinematics model of the robot is established and the kinematics simulation analysis of the robot is carried out.In order to realize the correct grabbing of the target object by the robot,the hand-eye calibration of the robot is studied,and the relationship between the robot and the base coordinate system is determined.Design the crawling software,establish a communication connection with the robot,and perform a grab experiment on the target object.The recognition effect of the robot on the target object has reached the expected...